作者位次 #4/6ACL 2026Findings· Findings of ACL
Idea First, Code Later: Disentangling Problem Solving from Code Generation in Evaluating LLM s for Competitive Programming
Sama Hadhoud · Alaa Elsetohy · Frederikus Hudi · Jan Christian Blaise Cruz · Steven Halim · Alham Fikri Aji
摘要
Large Language Models (LLMs) increasingly succeed on competitive programming problems, yet existing evaluations conflate algorithmic reasoning with code-level implementation. We argue that competitive programming is fundamentally a problem-solving task and propose centering natural-language edito...